Ann Thorac Surg 2000;69:1707-1710
© 2000 The Society of Thoracic Surgeons
Original articles: General thoracic
Perforation of Barrett’s ulcer: a challenge in esophageal surgery
Philippe G. Guillem, MDa,
Henri L. Porte, MDb,
Alain Saudemont, MDa,
Pierre A. Quandalle, MDa,
Alain J. Wurtz, MDb
a Service de Chirurgie Digestive et Générale, CAO, CAEB, Hôpital Huriez, France
b Service de Chirurgie Thoracique, Hôpital Calmette, CHetU Lille, Lille, France
Address reprint requests to Dr Porte, Service de Chirurgie Thoracique, Hôpital Calmette, CHetU Lille, 59037 Lille, France
e-mail: awurtz{at}chru-lille.fr
 |
Abstract
|
|---|
Background. Barrett’s ulcer, which develops within Barrett’s esophagus, is frequently responsible for bleeding. Perforation is a rare complication constituting a great challenge for diagnosis and management.
Methods. Three personal cases and 31 published reports of perforated Barrett’s ulcer were reviewed retrospectively. The site of perforation, clinical presentation, management, and outcome were assessed.
Results. The clinical presentation proved to be heterogeneous and was determined by the site of perforation: this was the pleural cavity (20% of cases), mediastinum (20%), left atrium (16.6%), tracheobronchial tract (13.3%), aorta (13.3%), pericardium (10%), or pulmonary vein (6.6%). Early esophagectomy and esophageal diversion-exclusion were the most frequent procedures, and overall mortality was 45%.
Conclusions. The poor prognosis of perforated Barrett’s ulcer should be improved by earlier diagnosis and adequate emergent operation. Although early esophagectomy constitutes the recommended procedure, esophageal diversion-exclusion, which allows control of both sepsis and bleeding, is also of interest.
|
Introduction
|
|---|
Barrett’s esophagus is defined as the result of the metaplastic process that replaces the squamous esophageal mucosa by a columnar epithelium [1]. It is thought to arise from the abnormal healing of a gastroesophageal reflux-mediated injury. This particular condition, found in approximately 10% of endoscopies performed for reflux symptomatology, is currently arousing more and more interest, because of its malignant potential [2]. However, the main subject of the initial report by Barrett in 1950 [3] was not the carcinogenic sequence but the peptic ulcer sometimes associated with the columnar epithelium. In such cases, the strict definition of Barrett’s ulcer (BU) is a peptic ulcer occurring within the metaplastic mucosa [3].
Few reports have been published about BU and its related complications. We report here three cases of perforated BU and review the English and French literature about perforated BU from 1950 to 1999, with special reference to clinical diagnosis and management.
|
Material and methods
|
|---|
During the 30 years from 1969 to 1999, 3 patients with esophageal perforation caused by BU were seen in our institution. All three were available for assessment of the site of their perforation and of its clinical presentation, management, and outcome. The following were excluded from the analysis: (1) patients with Barrett’s adenocarcinoma resulting in perforation or fistula formation, (2) patients with instrumental perforation during endoscopic examination of Barrett’s esophagus or during endoscopic dilatation for peptic stricture, (3) patients with esophageal tear after operation for reflux disease, and (4) patients with perforated esophageal ulcer without pathologic evidence of Barrett’s epithelium.
We also reviewed the 31 cases published between 1950 and 1999 that appear to fulfill the strict criteria for perforated BU (development of a benign peptic ulcer within Barrett’s epithelium). Many other cases were rejected because of uncertainty about the association of penetrating ulcer and pathologically proven Barrett’s esophagus, although some of them concerned ulcers that developed during peptic esophagitis.
Case 1
An 81-year-old woman was admitted for acute hematemesis. She had a 30-year history of chronic heart and respiratory failure because of severe kyphoscoliosis, but no symptoms of gastroesophageal reflux disease. On admission, the patient was conscious, with a systolic blood pressure of 90 mm Hg. A nasogastric probe produced dark brown material and bright red blood. Upper gastroesophageal endoscopy revealed a salmon-pink mucosa typical of Barrett’s esophagus, which had developed from the middle third of the esophagus, and in the distal esophagus a 4-cm-wide ulcer covered by fibrin and clots. Pathologic examination of biopsy specimens from the ulcer margin revealed a columnar epithelium with intestinal metaplasia. During the first 24 hours, the patient experienced massive hematemesis with hemodynamic shock and was therefore assigned to undergo emergent operation. Through laparotomy by means of the transhiatal approach, mediastinal dissection revealed an aortoesophageal fistula responsible for persistent active bleeding. Because of the patient’s poor general and hemodynamic condition, which ruled out thoracotomy, and dense mediastinal fibrosis, which prevented safe, direct control of the aortoesophageal fistula, a diversion-exclusion of the midesophagus was performed. The gastroesophageal junction and proximal part of the stomach were thus removed, after stapling of the lower part of the esophagus. Through a left cervicotomy, the cervical esophagus was divided, and the thoracic portion was excluded by manual suture. Substernal gastric pull-up was extended to the neck, where an esophagogastric anastomosis was performed. A feeding jejunostomy completed the intervention. The postoperative course was complicated by pulmonary sepsis with acute respiratory failure, and by a cervical anastomotic leak, which healed after 2 weeks with medical treatment. Oral intake was started on the 20th postoperative day. One year later, the patient was doing well and tolerated a normal diet. Thoracic computed tomographic scan showed only a slight dilatation of the excluded esophagus.
Case 2
A 64-year-old man presented with a 20-year history of hiatal hernia and gastroesophageal reflux disease, treated by Nissen fundoplicature, and a chronic respiratory failure caused by tobacco consumption and previous tuberculosis. On admission, the patient was febrile (38.3°C) and complained of slight dyspnea and episodic cough, which occurred when he was swallowing liquids. Physical examination was unremarkable, and laboratory studies showed a white blood cell count of 13,200/mL. Chest roentgenogram showed a right basilar infiltrate, which was confirmed by thoracic computed tomographic scan. Tracheobronchial endoscopy revealed a fistula foramen in the right lower lobe bronchus, and ingested methylene blue was seen flowing freely through the fistula from the esophagus. Esophageal endoscopy disclosed Barrett’s esophagus (of the fundic type, according to pathologic biopsy examination) in which a large ulcer had developed. Acute cough occurred during gas insufflation. Because of the patient’s poor respiratory condition, an exclusion-diversion of the esophagus was performed, with resection of the gastroesophageal junction and the proximal stomach through laparotomy by means of the transhiatal approach. This was followed by substernal gastric pull-up, cervical esophagogastric anastomosis through left cervicotomy, and feeding jejunostomy. The postoperative course was uneventful; oral intake was started on the 7th postoperative day, and the patient was discharged on the 21st day. One year later, clinical and radiologic examinations including chest roentgenogram and thoracic computed tomographic scan, were unremarkable.
Case 3
A 28-year-old man was admitted for epigastric pain, dyspnea, and sepsis. Two years previously, gastroesophageal endoscopy, performed for gastroesophageal reflux, had disclosed hiatal hernia, and an antacid treatment was instituted. On admission, his temperature was 38.6°C, and chest auscultation revealed a pleural friction rub. Laboratory tests showed a white blood cell count of 19,000/mL. The initial chest roentgenogram showed a right hydropneumothorax. A chest tube was inserted, and 500 mL of turbid fluid was drained. Ingested methylene blue was drained immediately, thus confirming the presence of an esophagopleural fistula. A water-soluble contrast swallow showed that the fistula extended from the esophagus to the right pleural cavity, 5 cm proximal to the gastroesophageal junction. Emergent laparotomy and right thoracotomy were then performed. The thoracic and abdominal portion of the esophagus were removed, the gastroesophageal junction was stapled, and the procedure was completed by left cervicotomy, cervical esophagostomy, and feeding jejunostomy. The postoperative course was complicated by sepsis, with multiple-organ failure leading to death on the 22nd postoperative day. Pathologic examination of the resected esophagus revealed the presence of Barrett’s ulcer within fundic and intestinal metaplasia.
|
Comment
|
|---|
Barrett’s ulcer, first described in 1950, is an uncommon complication of gastroesophageal reflux disease [3]. It concerns approximately 1% of patients with peptic esophagitis and 10% to 24% of those with Barrett’s esophagus [4–7]. Although perforation as a BU complication is mentioned in most reviews, it is in fact a rare evolutionary event. Accordingly, since the initial report in 1950 [3], we only found 31 reports concerning BU during the period up to 1999. To our review of these cases, we added the 3 patients treated in our department and described above, but excluded esophageal perforations caused by peptic esophagitis without BU. The clinical findings for the 34 cases are summarized in Table 1, and their therapeutic management is indicated in Table 2. Complications of BU, mostly bleeding events [3], are infrequent, and range from massive hematemesis [8], as in our first case, to occult bleeding with chronic anemia and iron deficiency [7]. Barrett’s ulcer has been shown to penetrate into all the mediastinal structures adjacent to the esophagus, such as the pleural cavity (20% of cases), left atrium (16.6%), tracheobronchial tract (13.3%), the aorta or its branches (13.3%), pericardium (10%), and pulmonary vein (6.6%). Occasionally, the perforation is contained within the mediastinum and does not penetrate a mediastinal organ (20%). Clinical features are consequently heterogeneous, depending on the site of the perforation, and sepsis is even inconstant (35%). However, the diagnosis of esophageal perforation is usually made on the basis of radiologic findings. Consequently, the fistula was only discovered at autopsy in six cases (26%). Chest roentgenogram may show direct signs of perforation, including hydropneumothorax, mediastinal emphysema, or pneumopericardium. Of interest is water-soluble contrast ingestion, which in most cases reveals the fistula and can therefore be recommended as the test of choice. Computed tomography scan can be a useful adjunct, as it sometimes shows pneumopericardium, esophageal wall thickness, mediastinal enlargement or abscess, or the fistula itself [9–11]. Esophageal endoscopy remains controversial, because theoretically, it may aggravate the clinical symptoms, because of air insufflation. However, it may affect the therapeutic management in identifying the underlying disease of the perforated esophagus, which has the endoscopic appearance of Barrett’s esophagus in 80% of the endoscopies performed for perforated BU.
Esophageal perforations are usually life-threatening. Their main prognostic factors include delayed diagnosis, a spontaneous mechanism, a thoracic site of perforation, and underlying esophageal disease [12]. In theory, perforation arising from BU fulfills all these criteria, a principle that has led to the recommendation of surgical management in most cases. Consequently, except for 2 patients whose very poor general condition prevented operation [13, 14] and 5 who died before any treatment, nonsurgical management was only the first choice treatment in three cases [11, 15]. In these cases, early diagnosis with minimal evidence that sepsis was confined to the mediastinum allowed this type of conservative management, which included broad-spectrum antibiotics, nasogastric suction, and supportive parenteral nutrition. This strategy conforms to the strict conventional criteria specified by several authors [16, 17] for nonsurgical management of esophageal perforations, ie, containment of the fistula within the mediastinum, accurate drainage, and confinement of clinical sepsis to a minimum. If one of these criteria is not fulfilled, an operation must be recommended as soon as the necessary resuscitation measures have been performed. In that case, however, the choice of procedure constitutes a great challenge. Except for 2 patients who underwent an antireflux procedure only, leading to ulcer healing in one [18], a direct approach to the perforated esophagus has been advocated by most authors, and all the methods previously described to treat esophageal perforation of causes other than BU that were used to manage perforated BU gave inconstant results (Table 2). They include (1) open drainage alone, through thoracotomy; (2) primary suture repair, generally in two layers, possibly supported by a local tissue flap [12]; (3) early esophagectomy; and (4) esophageal exclusion with synchronous or delayed gastric or colon interposition. Open drainage alone has been used in 2 patients with perforated BU. Exceptionally, primary suture repair was performed, and in our opinion should not be recommended, because when the esophagus is involved in severe esophagitis and Barrett’s esophagus, it is usually friable and inflamed, and holds suture poorly [19]. A recent review documenting the functional results of treatment for intrathoracic esophageal perforations demonstrated a significant need for repetitive dilatation and secondary esophagectomy in patients with gastroesophageal reflux disease undergoing primary suture [20]. Early esophagectomy therefore constitutes the treatment of choice, and was in fact used for 9 patients with perforated BU. It indeed allows complete resection of the abnormal esophagus, which led to the perforation. Furthermore, the presence of an obstructive lesion (peptic stricture) confirms the need for resection to relieve the obstruction for successful treatment of the perforation. Barrett’s esophagus is at least a well-known premalignant condition, and its removal is probably beneficial, especially in young patients like our third patient, who may have been exposed for a long time to the neoplastic evolution of the metaplasia. When esophagectomy cannot be performed because the patient’s poor general condition prevents thoracotomy, diversion-exclusion is an alternative procedure of interest, because it allows complete control of the septic process [17], as we demonstrated in our patients 1 and 2. For patient 1, who presented with an aortoesophageal fistula, mediastinal dissection for its direct closure would have been hazardous. We chose the never-reported option of esophageal diversion-exclusion. Spontaneous coagulation in the excluded part of the esophagus was probably responsible for drying up the bleeding and for the favorable outcome. Three previous reports of a fatal outcome after perforated BU extending into the aorta or one of its thoracic branches underlined the extreme seriousness of this condition, and the need for accurate emergent operation. Esophageal diversion-exclusion requires a synchronous or delayed substernal gastric pull-up. To facilitate early oral intake, we recommend immediate gastric interposition, which must be combined with prolonged high-dose antacid or proton-pump inhibitor therapy, as the latter was found to be beneficial for BU healing [5]. Whatever the surgical procedure, accurate drainage is required (eg, pericardiostomy for esophagopericardial fistula). Feeding jejunostomy should be combined with this drainage, to ensure early renutrition in debilitated patients and consequent tissue healing [12, 17]. For 31 of the 34 cases reviewed here, overall mortality was 45% (Table 2). This poor prognosis should be improved by early diagnosis and management. Only heightened awareness of the disease and suspicion of its presence can prevent delays or errors in diagnosis and allow appropriate decisions for its management.
|
References
|
|---|
-
Li H., Walsh T.N., O’Dowd G., Gillen P., Byrne P.J., Hennessy T.P. Mechanisms of columnar metaplasia and squamous regeneration in experimental Barrett’s esophagus. Surgery 1994;115:176-181.[Medline]
-
Pera M., Cameron A.J., Trastek V.F., Carpenter H.A., Zinsmeister A.R. Increasing incidence of adenocarcinoma of the esophagus and esophagogastric junction. Gastroenterology 1993;104:510-513.[Medline]
-
Barrett N.R. Chronic peptic ulcer of the esophagus and esophagitis. Br J Surg 1950;38:175-182.[Medline]
-
Ollyo J.B., Fontolliet C., Monnier P., et al. Pathogenic heterogeneity of Barrett’s ulcers. Apropos of 38 case reports. Schweiz Med Wochenschr 1989;119:747-751.[Medline]
-
Lee F.I., Isaacs P.E. Barrett’s ulcer. Am J Gastroenterol 1988;83:914-916.[Medline]
-
Komorowski R.A., Hogan W.J., Chausow D.D. Barrett’s ulcer. Am J Gastroenterol 1996;91:2310-2313.[Medline]
-
Williamson W.A., Ellis F.H., Jr, Gibb S.P., Aretz H.T. Barrett’s ulcer. J Thorac Cardiovasc Surg 1992;103:2-6.[Abstract]
-
Maiorana A., Stephens J.K., Little A.G. Barrett’s ulcer. Am Surg 1988;54:178-179.[Medline]
-
Backer C.L., LoCicero J.D., Hartz R.S., Donaldson J.S., Shields T. Computed tomography in patients with esophageal perforation. Chest 1990;98:1078-1080.[Abstract/Free Full Text]
-
Shah S., Saum K., Greenwald B.D., Krasna M.J., Sonett J.R. Esophagopericardial fistula arising from Barrett’s esophagus. Am J Gastroenterol 1998;93:465-467.[Medline]
-
Slim K., Elbaz V., Pezet D., Chipponi J. Traitement non chirurgical des perforations de l’oesophage thoracique. Presse Med 1996;25:154-156.
-
Michel L., Grillo H.C., Malt R.A. Operative and nonoperative management of esophageal perforations. Ann Surg 1981;194:57-63.[Medline]
-
Andersson R., Nilsson S. Perforated Barrett’s ulcer with esophago-pleural fistula. A case report. Acta Chir Scand 1985;151:495-496.[Medline]
-
Nousbaum J.B., Robaskiewicz M., Cauvin J.M., Mabille S., Gouerou H. Medical treatment of mediastinal perforation of Barrett’s ulcer (Traitement médical d’une perforation médiastinale d’un ulcère de Barrett). Presse Med 1989;18:228-229.
-
Altorki N.K., Skinner D.B., Segalin A., Stephens J.K., Ferguson M.K., Little A.G. Indications for esophagectomy in nonmalignant Barrett’s esophagus. Ann Thorac Surg 1990;49:724-726.[Abstract]
-
Cameron J.L., Kieffer R.F., Hendrix T.R., Mehigan D.G., Baker R.R. Selective nonoperative management of contained intrathoracic esophageal disruptions. Ann Thorac Surg 1979;27:404-408.[Abstract]
-
Goldstein L.A., Thompson W.R. Esophageal perforations. Am J Surg 1982;143:495-503.[Medline]
-
Diehl J.T., Thomas L., Bloom M.B., et al. Tracheoesophageal fistula associated with Barrett’s ulcer. Ann Thorac Surg 1988;45:449-450.[Abstract]
-
Katyal D., Jewell L.D., Yakimets W.W. Aorto-esophageal fistula secondary to benign Barrett’s ulcer. Can J Surg 1993;36:480-482.[Medline]
-
Iannettoni M.D., Vlessis A.A., Whyte R.I., Orringer M.B. Functional outcome after surgical treatment of esophageal perforation. Ann Thorac Surg 1997;64:1606-1609.[Abstract/Free Full Text]
Accepted for publication December 14, 1999.
This article has been cited by other articles:
|
|
|
|
 
J. J. Nigro, R. M. Bremner, C. B. Fuller, J. Theisen, Y. Ma, and V. A. Starnes
Perforating Barrett's ulcer resulting in a life-threatening esophagobronchial fistula
Ann. Thorac. Surg.,
January 1, 2002;
73(1):
302 - 304.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
